Appendix: illustrative supplier studies

The appendix is a case laboratory rather than a supplier survey. Every offer is fictional and deliberately controlled. The case names describe a mechanism to test, not a class of companies or a market average. Paired studies change an identified assumption while retaining the other supplied fields. This makes it possible to see why an ordering changes and which conclusion would be unjustified. An evaluator can reproduce the outputs through the companion workbench, then replace the fiction with supported comparable inputs before drawing a conclusion about a real procurement.

Read the pairs as separate questions. The short-service pair tests whether a larger acquisition survives a different authorised service duration. The energy pair tests a tariff difference for offers with different annual quantities. The downtime pair tests whether interruptions cause the supplied lost contribution or whether that loss is absent under the fictional recovery convention. The replacement pair tests a discrete service-life event at the horizon boundary. The residual pair tests a terminal credit whose reality remains a separate burden of proof. The rate pair tests how discounting changes the weight of later recurring cost relative to acquisition.

The printed input dictionary retains the full financial fields used by each bid. Common horizon, rate and electricity escalation appear beside each pair. A blank cost does not hide an omitted assumption: costs intentionally held absent are printed as canonical zero values. The detailed exported fixture remains the definitive full-precision input. The bound result tables identify each output bid by name because the engine sorts by lifecycle present cost. A bid that changes its position in the result has not changed its identity or input scope. This distinction matters when comparing controlled cases.

The service pair compares an acquisition-led offer with a service-led offer. The service-led fixture has a larger base acquisition, a longer supplied life and smaller uniform operating, energy and downtime quantities. The short requirement gives less time for those recurring differences to offset the premium. The longer requirement also brings a replacement into the acquisition-led alternative's cost. Both mechanisms are visible in the decomposed output. It would be wrong to explain the result only as a payback story because a discrete replacement and several recurring components contribute together.

The controlled change is the common horizon. Supplied life, annual duty, tariff, loss rate and acquisition remain unchanged. The longer case therefore answers a different service requirement with the same fictional offers. It does not prove that the buyer should choose the longer requirement or that the more durable offer will achieve its claimed life. The requirement must be authorised independently of the preferred bid. If continued need is uncertain, the buyer should retain both requirements and the consequence of committing to each rather than select the horizon that makes a favourite win.

The output reveals why a price premium can be defensible without proving that it is justified in practice. Its advantage rests on supplied operating and life differences that need relevant evidence. A supplier statement about longevity belongs beside its conditions, maintenance requirements and remedies. An energy difference belongs beside duty and measurement. A downtime difference belongs beside response and recoverability. If those assumptions are unsupported, the result is a useful description of the premise that would be needed, not an award recommendation or a prediction of realised savings.

The electricity pair separates physical quantity from the money assigned to it. The efficient offer uses less annual electricity in the supplied common service, but costs more to acquire. The lower-price case leaves the recurring value of the physical difference too small to recover that premium. The higher-price case gives the same quantity difference a larger monetary consequence. Neither case treats an increased tariff as an improvement in physical efficiency. The controlled change alters the price of a retained consumption difference, not the consumption difference itself.

Both offers have the same supplied life and operating amount, and the horizon excludes replacement. Downtime and terminal credit remain absent. Electricity escalation is also absent, so the pair isolates the input tariff without a second growth mechanism. The discounted electricity cost is still affected by the common rate and horizon. Those controls make the direction of the example easy to interpret. A real tariff can include demand charges, time variation or contracted blocks that the annual quantity-price multiplication does not represent; the evaluator should not assume that all electricity expense moves proportionately.

The practical evidence question follows the mechanism. If the premium depends on the lower annual quantity, a comparable duty and an appropriate measurement or verification arrangement matter more than an efficiency label. If the tariff is the decisive uncertainty, the buyer needs a defensible price convention or explicit alternatives. The pair demonstrates that the same technical offer can change its cost position when the economic boundary changes. It does not supply a future electricity price, establish avoided emissions or prove that a claimed operating quantity will be maintained in use.

The downtime pair holds interruption hours and all equipment inputs fixed while changing the supplied loss per hour. The deferred case gives interruptions no lost-contribution charge. It represents an intentionally restricted situation in which the relevant monetary loss is absent from this model. It does not assert that downtime is harmless: recovery labour, service penalties, inventory or disruption may still matter outside the fixture. The lost-contribution case assigns an economic consequence to the same hours. That change can justify a more expensive acquisition with fewer supplied interruptions.

The mechanism is incidence, not a different physical failure rate. The same assumed interruption quantity becomes costly when it displaces contribution that cannot be recovered within the relevant service. A revenue amount would not automatically be the appropriate loss rate because the interruption may also avoid variable cost or move demand to another period. The supplied rate must represent the buyer's relevant net consequence without duplicating a cost already in operating expense or a contractual charge. The pair's simple rate cannot settle that causal analysis, but it makes the dependence visible.

The actual decision record should retain the recovery route and its capacity. A statement that output can be recovered needs an executable schedule, available people and any extra cost. A statement that it is lost needs the causal connection to demand, contract or operating constraint. The pair provides a useful branch in that enquiry: if the loss mechanism is absent, the low acquisition can remain preferable; if the mechanism is present at the supplied magnitude, the lower-interruption offer can prevail. The computation does not establish which branch describes an actual operation.

The replacement pair removes discounting and holds recurring operating expense common. Its purpose is to isolate a discontinuity that smooth averages can conceal. In the shorter requirement, the short-life asset's next replacement falls exactly at the horizon and is excluded by the published convention. Extending the requirement places that replacement inside the horizon. The cost ordering then changes even though the base offers, life and annual operating amounts are unchanged. The mechanism is a required future purchase, not a marginal deterioration in efficiency or a newly assumed failure distribution.

The case is intentionally strict about terminal treatment. Both alternatives use no residual credit, even though they can have different remaining service outside the requirement. That restriction makes the replacement event transparent but limits how broadly the result can be interpreted. If retained service after the requirement has value, a defensible terminal convention may alter the comparison. The pair cannot establish equal lifetime utility simply because the study period is common. A committee should see both the discrete event and the no-credit boundary rather than assume the illustrative ordering is universally economic.

A practical requirement change can also create transition obligations beyond the replacement purchase. The buyer may need temporary service, recommissioning, disposal or a different maintenance arrangement. The workbench does not calculate those consequences from life automatically. They need supported explicit treatment in the comparison. The pair's exact boundary is therefore a diagnostic: it alerts the buyer that service extension changes the cost mechanism. It is not a recommendation to time a tender around a convenient replacement cut-off or a forecast that the equipment will last exactly as supplied.

The residual pair holds service life and every recurring amount common, so the larger acquisition begins with no operating advantage. The no-credit case exposes that premium directly. In the credit case, only the transferable offer receives a supplied terminal residual. Discounting moves that credit to the common base date and can reverse the lifecycle ordering. The pair therefore identifies an assumption that requires unusual attention: the decisive difference is an exit or retained-service value, rather than an efficiency or a lower running cost achieved during the operating horizon.

The label transferable is a fictional description, not evidence of a buyer, an enforceable sale or an accounting outcome. The terminal route would need a relevant condition assessment, authority to transfer, an accessible market or retained service plan, and net costs. Gross disposal proceeds and a net terminal credit are not interchangeable. Nor should a hypothetical sale be combined with continued-use value for the same asset unless the alternatives and incidences are clearly separated. The simple case makes none of those operational claims; it shows what the supplied credit does to present cost.

The paired result should change the evidence priority. If the case without a credit gives a different leader, a committee cannot treat the terminal amount as harmless detail. It can request support, retain a no-credit alternative, revise the requirement or choose another permissible response. The downloadable fixture makes that challenge reproducible: remove the credit and recalculate the entire output. The fact that a positive amount is accepted by the schema does not validate its economic basis. A supported residual remains a separate proposition from an exactly computed discounted residual.

The rate pair uses offers with different acquisition and operating costs while removing electricity, downtime and terminal value. Their supplied lives extend beyond the common horizon, so no replacement enters. The zero-rate case gives every annual operating amount equal weight. The higher-rate case reduces the present importance of later differences relative to the unchanged base acquisition. The resulting ordering demonstrates the role of discounting without presenting the higher rate as a market benchmark, an approved hurdle or a forecast of the buyer's financing cost.

The intervention changes only the common rate. It does not change the supplier price, service duration or operating amount. A decision record should therefore describe the result as a sensitivity to the valuation basis. It would be misleading to conclude that the lower-recurring-cost offer has become physically less efficient or that a lender has made it unfinanceable. Present-cost preference and liquidity are separate questions. A high rate can favour lower acquisition while a buyer's actual operating budget, service risk or policy constraint still prevents that alternative from being executable.

The evidence question is the rate's meaning. The team needs the relevant appraisal mandate, currency and inflation convention, and a clear distinction between analytical sensitivity and authorised central case. A rate should not be selected to reward a preferred offer. Nor should a universal public rate be copied into a different institution without checking its purpose. This pair demonstrates a mathematical boundary: recurring-cost advantage loses present weight as the rate changes. The actual decision needs an authorised basis and must preserve any separate non-cost or funding constraints.

The studies establish conditional differences rather than a universal supplier preference. A premium can survive through recurring savings, an avoided replacement or a supported terminal credit; it can fail when the requirement is shorter, the loss mechanism is absent or the valuation basis reduces future weight. Those are different causal routes. The evaluator should identify which route supports the actual proposal and seek evidence for that route. A generic assertion of better quality or lower lifetime cost is less useful than a specific premise whose consequence can be recomputed and challenged.

The pairs also show why an apparent robust leader can be a convention-dependent leader. A replacement boundary and a terminal credit are not smooth percentage sensitivities. A physical quantity and a tariff have different evidence sources. A recovered interruption and a lost one have different causal meanings. A common discount rate can preserve equal treatment while still being wrong for the appraisal mandate. The case laboratory gives each of those distinctions a durable numerical identity. It should encourage more faithful evidence and service definitions, rather than encourage the user to borrow the fictional quantities.

Choose the study pair by the disagreement rather than by the preferred supplier label. A changed service requirement belongs with the horizon pair; disputed purchased energy incidence belongs with the price pair; recoverable versus lost output belongs with the downtime pair. A replacement date belongs with the event pair, a net terminal route with the residual pair and a valuation mandate with the rate pair. The common lesson is a controlled comparison, but the evidence objects are different. Applying a tariff conclusion to a terminal promise would therefore miss the disputed mechanism even if both cause a ranking reversal. The cases remain complete fictional demonstrations of those distinctions.

A near-tie convention needs an explicit procurement purpose and unit; it should not be hidden in a numerical comparator. The small-distinct-lifecycle-costs study supplies two one-year acquisition-only bids with costs 1e-09 and 2e-09 USDk, with no continuing cost, terminal value or discounting. Their present costs are 1e-09 and 2e-09. The ordering remains distinct. Scientific display keeps these artificial values readable instead of rounding both to zero. Neither quantity is a market quotation.

The workbench orders computed lifecycle totals and warns of a numerical tie only when those totals are equal at its floating-point precision. The warning expressly leaves exact economic indifference unproved. Commercial materiality, scoring equivalence, bid uncertainty and the applicable tender treatment require an owned evaluation rule. A small calculated difference can be real within the illustration while being immaterial or unsupported in an actual procurement; that distinction belongs in the committee record.

A committee using a declared materiality margin should preserve the unequal totals and record the gap separately from its treatment. The evaluation record names the margin, its unit, the decision purpose, the approving owner and the uncertainty that matters. It can explain why a computed difference does not govern the choice without describing unequal arithmetic as a tie. Where the procedure requires a ranking, the owner must establish whether that separate treatment is permitted before using it. If the permission is unresolved, retain the computed ordering and defer the discretionary adjustment for review.